All gamma cameras include a lead collimator through which gamma rays are passed so that only those rays parallel to the slits in the collimator strike a scintillation crystal. The light of individual scintillations emanating from the scintillation crystal is not collimated but spreads out and travels through light tubes or fiberoptics to strike a plurality of photomultipliers which are usually arranged in a hexagonal array. The location of the point of scintillation origin is then obtained by algorithms based on a weighted average which analyzes all the individual signals from the photomultipliers. Specifically, the electrons or signals produced in the photomultipliers in response to the photons detected are essentially counted in pulses. Each pulse is formed into an intensity signal, z, which is correlated to the energy of the sensed photon(s) and a position signal, x,y, which is correlated to the point where the signal originated. The x-y and intensity signals are then corrected for energy, linearity and uniformity and, after a sufficient number of counts have been obtained, form specific pixels on a CRT (cathode ray tube) screen where the image of the radiated organ is produced.
Photomultipliers for gamma cameras are supplied with gain control boards. Typically, the photomultipliers are matched in the array to have about the same gain. Generally speaking, differences in gain between photomultipliers are accounted for by computer weighting of each photomultiplier during camera calibration vis-a-vis look-up tables stored in computer memory. In some instances the photomultipliers are purchased with an adjustable gain control effected by a potentiometer which is manually set or adjusted to a desired gain so that all the photomultipliers in the camera can have approximately the same gain.
More specifically, photomultipliers are typically selected with gain characteristics which are sized to produce substantially linear outputs for the photon energy levels which are detected. The photomultipliers are then calibrated by exposing the camera to a uniform known radiation source. See for example U.S. Pat. Nos. 4,866,615, 4,091,287 and 4,808,826. Typically a pinhole or slotted aperture lead mask is positioned in front of a reference radiation beam which produces a uniform radiation signal for all the photomultipliers. Manufacturing variations between photomultipliers cause variations in the photoanode output signal or variations in gain to occur among the photomultipliers. Heretofore, the industry has "adjusted" the variations by simply comparing the signals from all the photomultipliers and factoring them, mathematically, so that each photomultiplier's signal are mathematically adjusted to have the same gain as that photomultiplier which is the least sensitive or has the smallest gain in the photomultiplier array. The "gain" values for each photomultiplier are then stored in a "look-up" table within the camera's computer. When the isotope for which the camera has been calibrated is used in a patient, the "look-up" table factors each photomultiplier signal by the value stored in the table. It is appreciated of course that there is a separate look-up table for each isotope which the camera senses, and it is not uncommon for there to be as many as 27 or so look-up tables corresponding to the different radioactive isotopes used in the medical field today.
It is, or should be, obvious that the greater the gain signal difference between photomultipliers for any given isotope, the more significant the factoring becomes leading to the possibility of error. It also should be recognized that the more complicated the factoring becomes to extrapolate gain signals from exponential curves, a large amount of computer memory is required and the time for calibration is increased accordingly.
Using photomultipliers with manually adjustable gain control mechanisms does not resolve the problem discussed above. First, while the gain for each photomultiplier can be set to the approximate gain of one another when the camera is calibrated for one specific isotope, the adjustment is only approximate. It can never be precise. Secondly, while an approximate adjustment can be made for one isotope the fact that any one specific photomultiplier may not be linear for another isotope means the differences have to still be accounted for by automatically factoring the signals during calibration before the signals are stored in the look-up table. Thus, while manual gain control adjustments are helpful in that at least there is an attempt to remove any significant disparity between photomultipliers, they are not a solution to the problem.
Apart from variations in gain between photomultipliers which are supposedly resolved during calibration, calibration is also used to remove the noise inherently present in the photomultiplier. That is with the photomultiplier off, a signal is still produced at the anode which is termed noise. This signal is measured, stored and subtracted from the output signal produced during operation of the photomultiplier. Various techniques have been used to shut off the photomultiplier such as by tying the dynodes in the divider voltage resistor string together which maintains voltage in the voltage divider resistor string while shutting off the photomultiplier.
As is well known, the gain characteristics of the photomultipliers change in time producing errors. See for example U.S. Pat. Nos. 4,866,613 and 4,808,826 where the problem is discussed at some length. The solution followed by the industry as a whole has been to modify the compensating tables to compensate for the photomultiplier gain change and in this manner produce an accurate picture. However, because the compensation table itself is not linear, it is quite possible that any modification thereof as well as the initial table, can in turn produce error.
Insofar as controlling the gain of a photomultiplier, reference should be had to Photomultiplier Handbook, by Burle Industries, Inc, copyrighted 1980 (Printed 1989) Chapter 5 which discusses various gain circuits. The Handbook notes that while a separate voltage supply could be used for each dynode in the photomultiplier, a resistive voltage divider circuit is generally utilized for the dynodes and that the first dynode region should have a high cathode-to-first-dynode voltage for certain applications. Gain is then usually controlled by adjusting the overall or line voltage inputted to the voltage divider string. To avoid space charge effect, current in the voltage divider circuit should be ten times the anode current which, in the case of high gain gamma cameras, result in high power dissipations causing resistor heat which, in turn, affect the linearity of the photomultiplier's output signal. The Handbook also notes that certain dynodes can be tied so that the photomultiplier need not operate with all its stages and that where the overall voltage is not to be changed, it is possible to set the gain by setting the voltage of a single dynode. The frequently employed method to establish gain is to simply vary the overall voltage. With respect to automatic gain control circuits, reference can be had to U.S. Pat. Nos. 3,714,441 which utilizes a comparator circuit to adjust the line voltage to maintain the photomultiplier's gain at a desired value.